The market is finally waking up to the fact that AI spending is decelerating. But for those of us who trade crypto derivatives for a living, this isn't a macro footnote—it's a direct threat to the liquidity narrative propping up AI-related crypto assets.
Hook Over the past 12 months, the total market cap of AI-focused crypto tokens—Render (RNDR), Akash Network (AKT), Bittensor (TAO), and a dozen others—has surged over 400%, driven entirely by the expectation that enterprise AI capex would flow into decentralized compute. That expectation is now cracking. Goldman Sachs projects AI-related annualized spending could exceed $800 billion by end-2026; Morgan Stanley pushes that to nearly $3 trillion by 2028. But the critical detail is that 80% of that spending hasn't happened yet. And the rate of growth is slowing. Leverage doesn't care about your conviction—it cares about the marginal buyer.
Context The crypto market has historically been a leading indicator for speculative overhang. The same pattern emerges here: AI infrastructure spending is front-loaded, but revenue realization is back-loaded. The Bank for International Settlements (BIS) warned that the spending frenzy could morph into a long-term investment bust. In crypto terms, this is the equivalent of a protocol raising a massive treasury before proving product-market fit. The Aschenbrenner fund collapse—a $45B AI fund that imploded to ~$10B before Citadel took over—is the microcosm. That fund was levered on concentrated AI equity positions. Sound familiar? Crypto AI tokens are even more levered: no earnings, no cash flows, just narrative and speculation.
Core Let's examine the order flow. The top five hyperscalers (Microsoft, Amazon, Google, Meta, Apple) are expected to deploy over $1 trillion in 2025-2026 combined. That capital is flowing into NVIDIA GPUs, data centers, and storage. Storage stocks like Sandisk and Western Digital have rallied 396% and 145% respectively year-to-date. But this is a classic 'buy the rumor, sell the fact' setup. Once the capex guidance from these hyperscalers decelerates—and it will, because the incremental ROI on AI spend is diminishing—the entire chain will de-rate. In crypto, the equivalent is the GPU token narrative: projects like Render and Akash depend on hyperscaler overcapacity spilling into decentralized networks. If hyperscalers cut orders, GPU oversupply becomes a headwind for token prices, not a tailwind. We do not predict the storm; we short the rain.
Contrarian Angle The contrarian view is that slower AI spending could actually be bullish for decentralized compute. If hyperscalers pause their own infrastructure buildout, the demand for cost-effective, on-demand compute from Akash or Render could increase. But this is wishful thinking. The data shows that AI spending is concentrated in a handful of incumbents who will simply internalize the excess capacity rather than resell it to decentralized networks. The real contrarian insight is that the current AI capex is a 'defensive arms race'—companies invest not because they see a clear ROI, but because they fear falling behind. When that fear dissipates (as margins compress), the spending stops abruptly. Crypto AI tokens will be the first to feel the liquidity vacuum.
Takeaway I've been through the 2018 quiet audit of 0x Protocol, the DeFi leverage trap of 2020, and the NFT liquidity vacuum of 2021. The pattern is always the same: the market prices in exponential growth, but reality delivers logistic curves. The AI spending slowdown is not a forecast—it's already happening in the order book. If you're holding AI tokens, ask yourself: what is the marginal cost of the next GPU hour? If that cost is dropping faster than token inflation, you are the exit liquidity. Hedging is not fear; it is armor.